PROJECT SUMMARY/ABSTRACT Bariatric surgery is an important treatment option for morbidly obese patients who fail to lose weight through diet and exercise. Despite intervention, 20-50% of patients either fail to lose targeted amounts of weight or regain weight that was lost initially. Attempts at predicting the degree of weight loss have had only modest success and none have long term (>2 year) reliability. Moreover, there is a serious absence of research to predict weight loss beyond the 1st or 2nd year post-surgery and for outcomes other than weight loss including comorbidities common in the bariatric population. Our pilot data in 45 patients suggest that individual differences on pre-surgical neural activity measured with functional MRI (fMRI) reliably explains s 33% of the variance in weight loss up to 1 year post surgery, and over 50% of a multifaceted outcome measure, far outperforming many other indicators. These brain activation predictors implicate regions that closely conform to a theoretical model emphasizing both consummatory urges (a ?Now? neural circuit) vs. regulation of craving and self-control (a ?Later? circuit). Our central hypothesis is that individual differences in these neural pathways exert a powerful effect on the ability to sustain weight loss and achieve other key health outcomes. This project seeks to replicate and refine this model over a longer timeframe and to assess its predictive utility for key weight-related health outcomes. We propose to replicate the model derived from our fMRI pilot data predicting weight loss and secondarily to explore its predictive utility for changes in calorie intake, activity levels, liver fat, hemoglobin A1c, plasma lipids, blood pressure, and fasting glucose in a new, independent cohort of N=150 successively consenting, pre- surgical sleeve gastrectomy (SG) patients in study years 1-3. We will follow the pilot cohort for up to 7 years and the new cohort for 3 or more years to determine if predictors replicated in Aim 1 retain their long-term predictive power, particularly when supplemented with non-brain imaging variables and using a larger longitudinal dataset. We will use imaging and non-imaging data to develop multivariate statistical models incorporating energy balance, fMRI, and laboratory values with the variables described in Aim 1 to help to separate predictors vs. consequences of post-surgical outcomes. To help separate scan-to-scan variability from true post-surgical, trajectory-related brain changes, we will enroll N=20 obese subjects who will not undergo bariatric surgery, and are individually matched with our above SG subjects. Finally, in terms of translational potential, we will evaluate whether several related, non-fMRI cognitive tests that probe Now vs. Later functional domains to our MRI paradigms might have the potential to act as surrogate tests in clinical practice that help predict the likelihood of successful SG outcome during pre-surgical patient assessment. These studies are highly significant both in terms of understanding the role of neural patterns in weight regulation and in helping patients achieve bariatric population-specific health outcomes.